Introduction to the @cMosaic (cone mosaic) object
We illustrate basic methods for the cone mosaic class, @cMosaic. This class creates a cone mosaic arrays and computes the excitations from different optical images (OI).
We illustrate a near fovea @cMosaic object and compute a number of noisy response instances to a static stimulus with no fixational eye movements, so the computed responses consist of a single time point.
We also illustrate how to visualize the mosaic and to plot the computed cone excitations.
At the end we illustrate how get all the parameters from the @cMosaic constructor so you can reliably create the same @cMosaic. Or, you can create a second @cMosaic with the same parameters and use it to measure the response to a different stimulus.
See Also:
- t_cMosaicPlots
- t_cMosaicGenerate
- t_cMosaicSingleEyeMovementPath
- t_cMosaicMultipleEyeMovementPaths
- t_cMosaicEccDependentAbsorptionEfficacy
- t_cMosaicBenchMark
- t_cMosaicOffAxis
- t_cMosaicFromConeMosaicHex
For the simpler, rectangular mosaic, see coneMosaic.
Generate the ring rays stimulus
ieInit;
scene = sceneCreate('rings rays');
scene = sceneSet(scene, 'fov', 3);
Generate the mosaic
positionDegs = [1 0];
sizeDegs = [1 1];
cm = cMosaic(...
'size degs', sizeDegs, ... % SIZE: 1.0 degs (x) 0.5 degs (y)
'position degs', positionDegs, ... % ECC: (0,0)
'ecc varying cone blur', true ...
);
We have a small library of pre-computed mosaics. You can see which sizes and positions we have by typing
mosaicLoad('help')
You can load a stored mosaic and thus avoid the computation this way
cm = mosaicLoad(sizeDegs,positionDegs);
This is the standard human lens model. It is a good approximation over the central 10 deg (see Visual Encoding: Principles and Software, in the Circadian Rhythms book, Figure 6).
oi = oiCreate('human');
oi = oiCompute(oi,scene,'pad value','mean');
% If you want to try a specific OI for this visual field position, you can use
zCoeffDatabase = 'Artal2012';
eyeside = 'right';
pupilDiamMM = 3.0;
centerpsf = true;
wave = 400:10:700;
ecc = 1; % In degrees
subjectRank = 15; % [10, 40];
oi = ...
oiPosition(zCoeffDatabase, 'position',positionDegs, ...
'pupil diameter', pupilDiamMM, 'subject rank', subjectRank, ...
'wave',wave, ...
'eye side', eyeside,'center psf',centerpsf);
Visualize of the mosaic and the data
cm.visualize();
To compute a single noise free instance of the excitations
allE = cm.compute(oi);
cm.plot('excitations',allE);
To compute multiple noisy response instances of cone excitation response
instancesNum = 3;
[~, allNoisyE] = cm.compute(oi, 'nTrials', instancesNum);
for ii=1:instancesNum
cm.plot('excitations',allNoisyE(ii,1,:));
end
Make a line @regionOfInterest (ROI) and show it superimposed on the activations
% For help on ROIs, use this
% regionOfInterest('help');
roiLine = regionOfInterest('shape', 'line', ...
'from', [.5 0.2], 'to', [1.5,-0.2], ...
'thickness', 0.1);
% Show the ROI on top of the activations
cm.plot('roi',allE, 'roi',roiLine);
% Now choose a small circular ROI
roiCircle = regionOfInterest('shape','ellipse',...
'center',[1.3 0],...
'majorAxisDiameter',0.2,...
'minorAxisDiameter',0.2);
cm.plot('roi',allE, 'roi',roiCircle);
Retrieving the excitations within an ROI from all the excitations
roiE = cm.excitations('roi',roiCircle,'all excitations',allE);
mean(roiE)
ans = 86.8174
roiLE = cm.excitations('roi',roiCircle,...
'all excitations',allE, ...
'cone type','L');
mean(roiLE)
ans = 101.5606
% Fewer and lower excitations of S-cones.
[roiSE, roiSIdx] = cm.excitations('roi',roiCircle,...
'all excitations',allE, ...
'cone type','S');
mean(roiSE)
ans = 14.1930
Get the excitations and visualize in one call
[roiE, roiIdx, allE] = cm.excitations('roi',roiCircle,'visualize',true,'all excitations',allE);
Visualize responses - NC code. Thinking what to do
hFig = ieNewGraphWin;
activationRange = prctile(allNoisyE(:), [1 99]);
% Noise-free response
cm.visualize('figureHandle', hFig, ...
'axesHandle', gca, ...
'activation', allE, ...
'activation range', activationRange, ...
'vertical activation SliceEccentricity', -0.2, ...
'vertical activation ColorBar', true, ...
'plot title', 'noise-free response');
% Loop over noisy response instances
hFig = ieNewGraphWin;
for k = 1:instancesNum
cm.visualize('figureHandle', hFig, 'axesHandle', gca, ...
'activation', allNoisyE(k,:,:), ...
'activationRange', activationRange, ...
'verticalActivationSliceEccentricity', -0.2, ...
'verticalActivationColorBar', true, ...
'plotTitle', sprintf('noisy response instance (#%d)', k));
pause(1);
end
Making a reproducible cMosaic
For some calculations you would like to re-generate a repeatable, rather than random, cone mosaic. You can use the randomSeed slot for that.
% Here is the first mosaic. The second argument is the list of
% parameters used to create the mosaic.
[cm1, cm1P] = cMosaic(...
'sizeDegs', [1.0 1.0], ...% SIZE: 1.0 degs (x) 0.5 degs (y)
'positionDegs', [1 0], ...% ECC: (0,0)
'eccVaryingConeBlur', true, ...
'randomSeed', 12 ...
);
% Here is the second one, same random seed, so it matches
[cm2, cm2P] = cMosaic(...
'sizeDegs', [1.0 1.0], ...% SIZE: 1.0 degs (x) 0.5 degs (y)
'positionDegs', [1 0], ...% ECC: (0,0)
'eccVaryingConeBlur', true, ...
'randomSeed', 12 ...
);
% The parameters are equal because, well, we sent in the same
% parameters.
isequal(cm1P,cm2P)
% You can see the two mosaics are the same this way:
cm1.visualize;
cm2.visualize;
To create the mosaic from specifying the parameters, you can do this
Sometimes you just want to control all the parameters. So using the parameter list from above, we can create a mosaic this way.
[cm3, cm3P] = cMosaic(cm1P);
% Check that the parameters remained equal.
isequal(cm1P,cm3P)
% Have a look to see they are equal.
cm3.visualize;
You can change a parameter this way and see that it is different.
cm3P.randomSeed = 11;
cm4 = cMosaic(cm3P);
cm4.visualize;
You can also get the default set of parameters this way
% This creates a random mosaic.
cmP = cMosaicParams;
[cm, cmP1] = cMosaic(cmP);
cm.visualize;
% The random seed is returned - if you use the parameters again, you
% will get the same mosaic.
cmP1.randomSeed
ans = 859389